As we are approaching the end of the project, we are defining with CleanSky’s Topic Manager how the system will be evaluated for TRL5. First, to evaluate the system’s accuracy we will rely on recordings done in our cabin mockups with statistically sufficient samples for each of the following TTL (Taxi, Take-off, and Landing) cabin luggage requirement cases:

• Correct cabin luggage placement:

            -  Seat with passenger:

                     - Holding non-cabin luggage

                      - Nothing else

               - Seat without passenger:

                        - Non-cabin luggage on the seat

                        - Nothing else

• Incorrect cabin luggage placement:

                 -  Seat with passenger:

                         - Cabin luggage on the egress

                         -  Holding cabin luggage

                - Seat without passenger:

                         - Cabinluggage on the egress

                          - Cabinluggage on the seat

                 - Cabin luggage on the corridor

Where:

• Examples of cabin luggage: backpack, bottle, briefcase, camcorder, hat, laptop, shopping bag, suitcase, tote bag.

• Examples of non-cabin luggage: book/magazine, chocolate bar, fruit, music player, overcoat/jacket, sandwich, small purse, smartphone, tablet, tin can, umbrella, wallet.

All these elements need to be well-balanced not only to better analyse how the system performs in each case, but also to train it with the less bias possible.

Finally, by the end of the project, the system will be installed in CleanSky’s WP2.4 cabin demonstrator, including the hardware and software components to be able to test it on site.

‍